Archive for burns

in absurdum

Posted in Kids, pictures, Running, Travel with tags , , , , , , on July 3, 2023 by xi'an

“Il conduisait sans permis mais c’est un non sujet, vous avez vu comme moi que le gouvernement comptait faire passer l’âge du permis à 17 ans, alors bon…” J. Cambla, lawyer.

A few cars (and a bike!) were set on fire and a supermarket was looted just a block from home, on Thursday night, in one of the numerous riots that spread through French suburbs after a teenager was deadly shot while repeatedly trying to evade a police control. An underage driver with (obviously) neither driving licence nor insurance, in a car worth more than my yearly income. The most absurd [imho] aspect of these riots is that they destroy public and private property in the very place where these rioters live. As a result, the supermarket and the bakery [of bike fame] remain closed, the school where they sibling study may not reopen next Fall, their immediate neighbours have lost their car, and the local buses driving their parents to work no longer run…

unbiased MCMC with couplings [4pm, 26 Feb., Paris]

Posted in Books, pictures, Statistics, University life with tags , , , , , , , , , , , on February 24, 2020 by xi'an

On Wednesday, 26 February, Pierre Jacob (Havard U, currently visiting Paris-Dauphine) is giving a seminar on unbiased MCMC methods with couplings at AgroParisTech, bvd Claude Bernard, Paris 5ième, Room 32, at 4pm in the All about that Bayes seminar.

MCMC methods yield estimators that converge to integrals of interest in the limit of the number of iterations. This iterative asymptotic justification is not ideal; first, it stands at odds with current trends in computing hardware, with increasingly parallel architectures; secondly, the choice of “burn-in” or “warm-up” is arduous. This talk will describe recently proposed estimators that are unbiased for the expectations of interest while having a finite computing cost and a finite variance. They can thus be generated independently in parallel and averaged over. The method also provides practical upper bounds on the distance (e.g. total variation) between the marginal distribution of the chain at a finite step and its invariant distribution. The key idea is to generate “faithful” couplings of Markov chains, whereby pairs of chains coalesce after a random number of iterations. This talk will provide an overview of this line of research.

art brut [prolégomènes]

Posted in pictures with tags , , on June 10, 2012 by xi'an